The soluble proteins of the bovine cornea
William S. Holt and Jin H. Kinoshita
The soluble proteins of calf cornea, both epithelial and stromal, were fractionated by Sephadex
G-200 gel filtration followed by DEAE-cellulose ion-exchange chromatography. Each fraction
was further analyzed using a variety of techniques, including disc electrophoresis in polyacrylamide gels, Ouchterlony immunoprecipitation, and immunoelectrophoresis. From the corneal epithelium some 20 or more distinct protein zones are demonstrated by electrophoresis.
A single protein appears to comprise about 40 per cent of the epithelial proteins and a
significant part of the stroma as well. The epithelium is devoid of serum proteins. The stroma,
however, contains at least five serum proteins, albumin, 7S-gamma globulin, transferrin, a
lipoprotein, and an unidentified protein of beta electrophoretic mobility. These serum proteins
constitute over half of the soluble stromal proteins. Of the remainder, there is one stromal
protein identical with the principal epithelial protein, a few are glycoproteins, and none are
soluble collagen.
A,
rum albumin as a model for calculating
the diffusional properties of the cornea
and demonstrated that a protein with the
electrophoretic mobility of albumin was
present in rabbit corneas. Certain studies 8 - 9
suggested that soluble proteins were involved in the immunologic sensitization of
the recipients of corneal grafts. The present study was undertaken to gain knowledge of the soluble corneal proteins as a
basis for possible further study of corneal
transplantation antigens.
.lthough the insoluble components of
the cornea, such as collagen and the sulfated mucopolysaccharides, have been investigated in detail, very few reports exist
describing the soluble protein components.1
A few report the relative amounts of corneal protein which are soluble,2'3 one suggests the presence of certain serum proteins,3 two utilize the fluorescent antibody
technique to demonstrate the presence of
serum albumin and immunoglobulins,4-5
and one older paper studied soluble glycoproteins.0 Maurice and Watson7 used se-
Methods
The corneal epithelium of calf eyes was scraped
off with a scalpel blade, homogenized in 0.01 M
phosphate buffer, pH 7.6 usually was used. Following centrifugation at 15,000 x g for 30 minutes,
the supernatant fluid was dialyzed overnight
against the same buffer. The resulting solution contained about 15 mg. per milliliter of soluble epithelial proteins. The stroma was then excised,
minced with scissors, and thoroughly homogenized
in a small Waring blender. The resulting fine slurry
was stirred overnight in the cold, centrifuged
30 minutes at 15,000 g, and the supernatant solution decanted and saved. The pellet resuspended
From the Howe Laboratory of Ophthalmology,
Harvard Medical School, Massachusetts Eye and
Ear Infirmary, Boston, Mass.
This investigation was supported by Public Health
Service Training Grant No. 5T01EY00018 from
the National Eye Institute.
Manuscript submitted July 17, 1972; manuscript
accepted Nov. 6, 1972.
Reprint requests: Dr. Jin H. Kinoshita, Bldg. 6,
Rm. 222, National Eye Institute, Bethesda, Md.
20014.
114
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Volume 12
Number 2
in buffer, stirred overnight, and recentrifuged.
The two supernatants were combined, dialyzed for
24 hours against a large excess of buffer, and concentrated to 10 to 15 mg. per milliliter by placing
the solution, contained in dialysis tubing, in dry
Sephadex G-200 power overnight. Protein concentrations were estimated by the optical density10
and Lowry11 methods.
Sephadex G-200 gel filtration was performed in
1.7 by 50 cm. columns, approximately 75 mg. of
stromal or epithelial soluble proteins being
applied. DEAE-cellulose chromatography was performed on each of the main Sephadex fractions in
1.7 by 50 cm. columns previously equilibrated
with 0.01 M phosphate, pH 7.6. The first DEAE
fraction was eluted with a 0.01 M phosphate
wash, the remaining fractions being eluted at
various points along a concentration gradient
achieved with a Marriott flask arrangement in
which 175 ml. of 0.15 M phosphate, pH 7.6, was
added drop by drop with mixing to 120 ml. of
0.01 M phosphate. This gradually raised the eluting
molarity for 0.01 to 0.12 M over 175 ml. Next, 250
ml. of 0.2 M phosphate was similarly added,
gradually raising the eluting molarity from 0.12 to
0.19. All the proteins were eluted well before this
point.
Disc electrophoresis was performed on each
Sephadex and DEAE fraction by the method of
Davis,12 except that the separating gel was 6.5
per cent acrylamide. Densitometry of the stained
gels was done with a disc electrophoresis densitometer attachment to the Gilford recording spectrophotometer (Gilford Institute, Oberlin, Ohio).
The mobility of a protein zone relative to the
bromphenol blue tracking dye (the Rf value) was
calculated from these densitometry traces and
from direct measurements of the gels. In certain
cases, the gels were stained for glycoproteins by
the modified periodic acid-SchifF (PAS) method of
Zacharius and co-workers.13 Staining of the gels
for lipoproteihs and serum tranferrin was done
by the methods of Beaton, Selby, and Wright,14
and for haptoglobin by the method of Smithies.15
Nucleic acid content of the fractions was estimated
by the orcinol method.10 Hydroxyproline was
estimated by the method of Neuman and Logan,17
with protein hydrolysis and removal of chromogens
by the methods of Prockop and Udenfriend.18
Rabbit antiserum to soluble calf corneal proteins
was produced in 5 to 8 pound albino rabbits by
a series of 8 immunizations, the first 5 at weekly
intervals, the last 3 at two week intervals. At each
immunization, a total of 10 mg. soluble corneal
proteins from homogenates of the whole cornea
mixed 1:1 in Freund's complete adjuvant was
given intradermally in four separated sites. The
animals were bled by marginal ear vein initially
and 3 days after each immunization. The sera
were assayed for antibody activity by the Ouch-
Soluble proteins of bovine cornea 115
terlony method.19 No improvement in titer was
noted after the sixth immunization. Commercial
antisera used were rabbit antibovine serum proteins (GIBCO, Grand Island, N. Y.), rabbit antibovine gamma globulin, and rabbit antibovine
serum albumin (Behringwerke AG, obtained
through Certified Blood Donor Service, Inc.,
Woodbury, N. Y.).
Double-diffusion immunoprecipitation was performed by a micromodification of the basic
Ouchterlony technique,19 using 1.5 per cent agar
in 0.05 M phosphate, pH 7.6. Immunoelectrophoresis was basically by the micromethod of
Scheidegger,20 using 2 per cent agar in Veronal
buffer, pH 8.2, ionic strength 0.05.
Results
The average wet weights and yields of
proteins extracted from calf corneas are
presented in Table I. Assuming the dry
weight of the stroma to be 25 per cent
of the wet weight,3 soluble protein represents 11.5 per cent of the stromal dry
weight. This and other figures in Table
I are in reasonable agreement with data
of previous studies.13 The soluble proteins
were fractionated according to the general
scheme outlined in Fig. 1. In designating
the protein fractions, "E" denotes that the
fraction is of epithelial origin, "S" denotes
stromal, "G" denotes a Sephadex G-200
fraction, and "D" one from DEAE.
Gel filtration on Sephadex G-200 of
either epithelial or stromal soluble proteins
yielded two fractions, as shown in Fig.
2. Blue dextran used as a marker eluted
from the column just ahead of the first
stromal fraction, S-Gl, suggesting it is
composed of proteins retarded slightly by
the gel. Peak S-Gl exhibited a "knee" on
its descending arm. Disc electrophoresis
(Fig. 5, gels 1 and 2) revealed that this
knee represents the addition of new protein not present in the earlier part of peak
S-GL The relative amounts of protein eluting in each fraction are given in Tables
II and III. Epithelial fraction E-Gl exhibited greater absorption at 260 than at
280 mfi, suggesting material containing
nucleic acid. The ratio of nucleic acid to
protein, each in milligrams per milliliter,
was calculated for each Sephadex fraction. For E-Gl it was 1.2, whereas for E-G2
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Investigative Ophthalmology
Febmary 1973
116 Holt and Kinoshita
Table I. Extraction of cornea soluble proteins
Wet weight
Epithelium
Stroma
Entire cornea
mg.
14
211
225
Per cent of total
6.2
93.8
100
Table II. Approximate yields of soluble
epithelial proteins in the Sephadex and
DEAE fractions
Per cent of
soluble
Per cent of
Protein per epithelial
fraction
Fraction cornea (mg.) protein
E-G2
E-Gl
E-G2
E-G2-D1
E-G2-D2
E-G2-D3
E-G2-D4
and 5
0.5
1.5
0.26
0.78
0.10
0.36
25
75
13
39
5
18
Per cent
soluble protein
(mg. sol. prot.)
Per cent of total (mg. wet wt.)
25
14.1
75
2.8
100
3.5
Soluble protein
100
17
52
7
24
it was only 0.11, for S-Gl, 0.17, and for
S-G2, 0.09, thus confirming this impression.
Disc electrophoresis of E-Gl is shown in
Fig. 4.
Further purification of each epithelial
and stromal Sephadex fraction was attempted on DEAE-cellulose columns. Fraction E-Gl did not yield discrete peaks
and no further analysis was pursued. Stromal Sephadex fraction S-Gl was separated
by DEAE into several very small peaks
and a single large one which eluted at a
phosphate buffer concentration of about
0.08 M. Disc electrophoresis (Fig. 5, gel
3) revealed that the large 0.08 M peak
contained the two principal S-Gl protein
zones in large amounts and relatively free
of contaminants. Accordingly, tubes eluting
from 0.075 to 0.0S5 M phosphate were
pooled, labeled as fraction S-G1-D3, and
used in further studies.
The Sephadex fractions E-G2 and S-G2
were also subjected to DEAE chromatography, with the results illustrated in Fig.
3. The DEAE fractions are labeled Dl
through D5 with a prefix E-G2 or S-G2
mg.
1.98
5.89
7.87
Table III. Approximate yields of soluble
stromal proteins in the Sephadex and
DEAE fractions
Fraction
S-Gl
7s-gamma
globulin
Rf .08
serum
protein
other S-Gl
S-G2
S-G2-D1
S-G2-D2
S-G2-D3
S-G2-D4 & 5
Albumin
Rf. 58
Other
Per cent
Protein per of soluble
cornea
stromal
protein
Per cent
of fraction
S-Gl or
S-G2
3.2
55
100
1.8
0.4
31
56
14
1.0
2.7
0.29
0.83
0.53
1.0
0.7
0.15
0.15
7
17
45
5
14
9
17
(12)
(2.5)
(2.5)
30
100
11
31
20
38
Designations of fractions are as in the text and in Fig. 1.
Bases for calculations are milligram yields from Sephadex
or DEAE columns and triangulation estimates of areas
under densitometry traces of disc electrophoresis gels.
denoting the Sephadex fraction of origin.
Peaks E-G2-D1 and S-G2-D1 were eluted
with the 0.01 M phosphate wash. Subsequent E-G2 peaks were eluted at 0.045,
0.075, 0.10, and 0.17 M phosphate. The
four subsequent S-G2 peaks were eluted
at approximately 0.050, 0.080, 0.11, and
0.18 M phosphate. The relative amounts
of protein in each fraction are given in
Tables II and III. Each of these fractions
was collected, concentrated, and further
analyzed by disc electrophoresis.
Disc electrophoresis was performed on
each of the Sephadex and DEAE fractions, as illustrated in Figs. 4 through 7.
The protein zones resolved by electrophoresis can be categorized in terms of their
Rf values, an approximate scale of which
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Soluble proteins of bovine cornea 117
Volume 12
Number 2
EPITHELIAL SOLUBLE
PROTEIN
I.Sephadex chromatography
EG2
EG1
II.DEAE chromatography
i
Dl
D3
D2
D4
. Disc electrophoresis &
immunochemical studies
Principal epithelial
protein Rf .32
No fraction contained serum proteins
STROMAL SOLUBLE
PROTEIN
I. Sephadex chromatography
SG2
II. DEAE chromatography
D3
D2&
2a
D3
D4
III. Disc electrophoresis &
immunochemical studies
-Gamma gobulin
-Serum lipoprotein
-Slow mobility
stromal proteins
-Epith.-stromal
protein Rf.32
-Transferrin
-Stromal
glycoproteins
-albumin
-repeating
stromql
proteins
Fig. 1 A and B. Flow diagrams showing overall separation scheme employed for fractionating
the soluble stromal and epithelial proteins of the calf cornea. Each fraction is designated by
a set of letters and numbers as described in detail in the text and subsequent figures.
appears beside the gels in these figures.
In Fig. 4, the first epithelial Sephadex
fraction, E-Gl, reveals only a few faint
zones. The second Sephadex fraction,
E-G2, is more complex, showing at least
11 bands clearly. Of the five DEAE fractions into which E-G2 is resolved most
contain several protein zones. The fraction
E-G2-D2, however, which makes up 52
per cent of E-G2, appears to contain a
single protein, labeled "e," with Rf 0.32.
Fractions E-G2-D4 and -D5 had similar
electrophoretic appearances, so only E-G2D4 is illustrated. Disc electrophoresis of
stromal Sephadex fraction S-Gl is shown
in gels 1 and 2 of Fig. 5. Gel 3 in the
figure is the result of electrophoresis of the
DEAE fraction S-G1-D3, described earlier
as coming from that portion of S-Gl eluting off DEAE at 0.075 to 0.085 M phosphate. It contains a single sharp zone,
labeled "s," and a broader heterogeneous
zone "g," the identities of which will be
discussed later.
The second stromal Sephadex fraction,
S-G2, resolved into a number of protein
zones on disc electrophoresis, as seen in
gels 4 and 5 of Fig. 5. The more concentrated, tube 5, shows certain minor components to better advantage, while the
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Investigative Ophthalmology
February 1973
118 Holt and Kinoshita
4.0
STROMAL
EPITHELIAL
Z 2.O-
10
20
FRACTION
30
40
Fig. 2. Elution of soluble stromal and epithelial proteins from Sephadex G200. Two major
fractions are eluted from both stroma and epithelium. These are labeled as indicated by the
bars under each peak and as outlined in Fig. 1 and the text. In the curves depicted, 100 mg.
of stromal and 60 mg. of epithelial proteins were applied to the respective columns.
less concentrated, tube 4, shows the major
zones better. DEAE chromatography of
S-G2 gave the 5 fractions described earlier,
the electrophoretic analysis of which is
presented in Fig. 6. Of the many protein
zones revealed in each fraction certain ones
were studied further. Fraction S-G2-D2
contains 2 proteins, labeled "e" and "t,"
migrating closely together. These are also
shown in the gel D2(a), which is taken
from an intermediate fraction midway between S-G2-D2 and S-G2-D3. The "e"
zone will be shown later by immunoprecipitation to be the same protein as the
principal epithelial protein described
earlier in fraction E-G2-D2. Fraction SG2-D3 appears to have traces of the "e"
and "t" zones as well as a heavily staining,
more rapid component, Rf about 0.65.
DEAE factions S-G2-D4 and S-G2-D5 had
similar electrophoretic appearances, and so
only -D4 is illustrated. The heaviest -D4
zone, Rf 0.76, labeled "a" in Fig. 6, will
later be shown to be serum albumin. Migrating behind albumin is a second heavily
stained zone (labeled "c") and a series
of regularly spaced zones of progressively
decreasing concentration and mobility,
labeled by dots. The appearance of these
suggests a series of aggregates of a basic
structural unit such as is seen in the human
serum haptoglobins.21
Each of these electrophoresis gels was
stained by the PAS method for glycoproteins,18 and those with positive results are
shown in Fig. 7. Although no glycoproteins
were demonstrated by this method in any
epithelial fractions the stroma contained
several. Stromal fraction S-Gl exhibited
2, the sharp "s" zone and the diffuse "g"
protein. The "t" zones of fractions S-G2-D2
and S-G2-D2(a) stained clearly, as did a
diffuse zone just above. The most prominent stromal glycoprotein is in fraction
S-G2-D3, labeled "p."
The production of a potent anticorneal
protein antiserum proved difficult, even
after repeated immunizations. Ours reacted
with an unfractionated homogenate of soluble corneal proteins to give 4 precipitin
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Volume 12
Number 2
Soluble proteins of bovine cornea 119
1.0-
STROMAL G2
PITHEllAl G2
Fig. 3. DEAE-cellulose chromatography of the two Sephadex fractions, epithelial fraction E-G2
(squares), and stromal fraction S-G2 (circles). The peaks are labeled as described in the flow
diagrams of Fig. 1 and in the text. The dashed line represents eluting buffer molarity. Each of
these peaks was further studied by disc electrophoresis and immunoprecipitation.
lines. Two of these were removed after
absorption against calf serum, leaving two
corneal specific antigens. One of these
corneal specific antigens proved to be present only in stromal fraction S-G2-D4, as
shown in Fig. 8, A. The other was present
in the epithelial fraction E-G2-D2, and
must therefore correspond to the Rf 0.32
("e") electrophoretic zone (see Fig. 4).
This same antigen is also present in the
stromal fraction S-G2-D2 and in trace
amounts in S-G2-D1 and -D3, as is illustreated in Fig. 8, B. This antigen probably
corresponds to the "e" zone in Fig. 6. This
"e" zone represents about 39 per cent of
the soluble epithelial protein and a significant part of the stromal fraction, S-G2.
The antiserum to calf serum proteins
delineated no serum proteins in any epithelial fraction but revealed several in
the stroma, as is illustrated by immunoelectrophoresis in Figs. 9, A and B. The
most cathodic precipitin arc in S-Gl, labeled "g," is shown in Fig. 9, B to be gamma globulin. Although the antiserum used
is polyspecific for all gamma globulins, the
antigen delineated here is most likely a
7S globulin, as 19S globulins are unable to
diffuse into the cornea.7 This S-Gl antigen probably corresponds to the electrophoretic zone "g" of Fig. 5 as this resembles
the known disc electrophoretic appearance
and mobility of 7S gamma globulin.12-21
The positive PAS staining of this electrophoretic zone (Fig. 7) is supportive evidence, as gamma globulin is a glycoprotein.22 Densitometry showed this zone to
make up 56 per cent of fraction S-Gl
(Table III), which means that about 31
per cent of the soluble stromal protein
is gamma globulin, this being about 1.8
mg. per calf cornea.
A second serum antigen in fraction S-Gl
is represented by the arc labeled "s" in
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
r
Investigative
120 Holt and Kinoshita
Ophthalmology
Fcbruaty 1973
5 <**
Gl
G2
Dl
D2
D4
-to
Fig. 4. Disc electrophoresis of soluble epithelial proteins. Gels; left to right, are: unfractionated
epithelial proteins (E), Sephadex fractions E-Gl and E-G2, and DEAE fractions E-G2-D1,
-D2, -D3, -D4 {see Fig. 1). Migration is from top to bottom. The Rf scale is approximate and
meant only as a general guide. Labels: e-the "principal" epithelial protein, also found in
stromal fraction S-G2-D2 (see Fig. 6).
.Or
1.0L
1
Fig. 5. Disc electrophoresis of stromal proteins. Gel 1: fraction S-Gl from ascending limb
of elution curve S-Gl; gel 2: S-Gl from descending limb showing addition of a new protein
Rf 0.59; gel 3: that portion of S-Gl eluting at approximately 0.08 M phosphate from DEAE
and labeled S-G1-D3; gel 4: S-G2, 65 fig sample applied, gel 5: S-G2, 100 ^g applied.
Labels: g-7S gamma globulin, s-the Rf 0.08 serum lipoprotein, and a-albumin.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Volume 12
Number 2
Soluble proteins of bovine cornea
.0
1.0-
Dl
D2
D2(a) j
D3
Fig. 6. Disc electrophoresis of stromal S-G2-DEAE fractions. Tubes are designated, from left
to right, fractions S-G2, S-G2-D1, S-G2-D2, S-G2-D2 (a), S-G2-D3, and S-G2-D4. Labels:
e-the Rf 0.32 principal epithelial protein, also present in stroma, t-transferrin, a-albumin,
c-the Rf 0.57 stromal protein, and dots-the repeating stromal protein, as described in the text.
Fig. 7. Glycoprotein staining of selected stromal fractions. Gels la, 2a, and 3a are stained with
Coumassie blue, a protein stain, while lb, 2b, and 3b are stained by the PAS method13 for
glycoproteins. Gels la and lb: fraction S-Gl, gels 2a and 2b: S-G2-D2, and gels 3a and 3b:
S-G2-D3. Labels: g-gamma globulin, s-the Rf 0.08 serum lipoprotein, t-transferrin, and p-the
principal stromal glycoprotein.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
121
122 Holt and Kinoshita
Investigative Ophthalmology
February 1973
Fig. 8. Ouchterlony precipitin assays of soluble calf corneal protein fractions. A. Center well:
stromal fraction S-G2-D4. Peripheral wells: 1, 3, 5 anticalf corneal proteins; wells 2, 6 anticalf
serum proteins; well 4 anticalf gamma globulin. This shows a serum and a corneal specific
antigen in S-G2-D4. B. Center well: anticalf corneal proteins. Peripheral wells: ( 1 ) E-G2-D1,
( 2 ) S-G2-D1, ( 3 ) E-G2-D3, ( 4 ) S-G2-D3, ( 5 ) S-G2-D2, ( 6 ) E-G2-D2. The corneal specific
antigen in epithelial fraction E-G2-D2 is immunochemically identical to the one in S-G2-D2
and -D3, and probably S-G1-D1. C. Center well: anticalf albumin. Peripheral wells: ( 1 ) unfractionated stromal proteins, (2) S-G1-D3, ( 3 ) S-G2-D1, ( 4 ) S-G2-D2, ( 5 ) S-G2-D3, and
( 6 ) S-G2-D4. Albumin is found in S-G2-D4. D. Center well: S-G2-D2. Peripheral wells: 1, 3,
5 anticalf corneal protein; 2, 4, 6 anticalf serum proteins. S-G2-D2 contains a serum and a
nonserum antigen.
Figs. 9, A and B, and corresponds to the
other major electrophoretic zone in S-Gl,
labeled "s" in Fig. 5. This disc electrophoretic zone gives a positive glycoprotein
stain (Fig. 7) and was the only zone to
stain positively with oil Red-O, a lipid
stain.14 By densitometry, this zone makes
up 14 per cent of S-Gl, or about 7.7 per
cent of the soluble stromal proteins. The
third serum antigen in S-Gl, labeled "u"
in Fig. 9, A, was not seen when smaller
concentrations of sample were applied, as
in Fig. 9, B, and therefore probably represents a serum protein present in small
amounts. Benzidine staining for hepatoglobin15 was negative and the identity of
this protein was not established.
The S-G2 antigen with alpha mobility,
labeled "a" in Fig. 9, A, appears to be serum
albumin. In Fig. 8, A, fraction S-G2-D4
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Volume 12
Number 2
Soluble proteins of bovine cornea 123
Fig. 9. Immunoelectrophoresis of stromal protein Sephadex fractions. A. Upper well: Fraction
S-Gl. Lower well: S-G2. Trough: antibody to calf serum proteins. B. Center well: Fraction
S-Gl. Top trough: anticalf serum proteins. Lower trough: anticalf gamma globulins. Three
serum antigens are in S-Gl and 3 in S-G2. A smaller amount of S-Gl was applied in Fig. 9, B
than in 9, A, and only two serum components are seen. Labels: g-gamma globulin, s-lipoprotein, u-unidentified, Mransferrin, and a-albumin.
was shown to contain one serum and one
corneal specific antigen, and Fig. 8, C illustrates that the serum antigen is albumin.
This figure also shows the lack of albumin
in the other S-G2 DEAE fractions. The
"a" electrophoretic zone of fraction S-G2D4 in Fig. 6 is thus almost certainly albumin, exhibiting the expected rapid mobility
and characteristic ability to bind the bromphenol blue tracking dye during electrophoresis.12 By densitometry, this "a" zone
makes up 70 per cent of S-G2-D4 and -D5
(Table III) and albumin thus represents
about 12 per cent of the soluble stromal
protein. The cornea-specific antigen demonstrated in Fig. 8A must correspond to
one of the other electrophoretic zones of
S-G2-D4, but it is uncertain which one.
The precipitin arc with beta mobility in
Fig. 9, A, labeled "t," was found to be
localized to stromal fraction S-G2-D2 by
Ouchterlony analysis. This fraction contains
only two proteins by electrophoresis (Fig.
6) and two antigens by Ouchterlony analysis (Fig. 8, B), Of these two, the corneal
specific antigen has already been related
to the Rf 0.32 "e" protein electrophoretic
zone. Thus, the serum antigen must correspond to the series of two to three closely spaced electrophoretic zones labeled "t"
in Figs. 6 and 7, the mobility and appearance of which closely resemble that of
bovine transferring4 an iron-binding serum
beta glycoprotein.-'2'23 Specific staining of
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Investigative Ophthalmology
February 1973
124 Holt and Kinoshita
this fraction for transferrin by the ironNitroso R salt method14 gave positive results with these Rf 0.36 to 0.40 zones. Transferrin is a glycoprotein22; thus the positive
PAS staining of these zones in Fig. 7 is
confirmatory. The final precipitin arc in
fraction S-G2, labeled "g" in Fig. 9, A, bottom, was shown by Ouchterlony studies not
illustrated to represent traces of gamma
globulin in fraction S-G2-D3.
None of the protein fractions contained
collagen in amounts that could be detected
by the assays used.17'18 The negative results are consistent with those of Polatnick, Tessa, and Katzin2 who could extract no soluble collagen at neutral pH.
Discussion
This study was undertaken to survey the
soluble comeal proteins of the calf using
standard methods. Even though complete
purification of each of the main corneal
proteins was not achieved, the existence of
a large number of discrete proteins has
been demonstrated, and the use of various
methods has permitted several deductions
and conclusions to be made about them.
The quantitative considerations presented
in Tables II and III must be considered
approximate, especially those based on
densitometry of the electrophoretic gels,
as this method clearly distinguished the
major zones but not the minor and closely
spaced ones. The data are presented because few other quantitative estimates of
the various soluble corneal proteins are
available.
The corneal epithelium is shown to
contain a host of discrete proteins, mostly
in small quantities, and presumably, mostly of intracellular origin. In the light of
this, it is interesting that one protein fraction, with Rf 0.32 on disc electrophoresis,
makes up about 40 per cent of epithelial
soluble proteins. A possible importance
of this protein is that it is also found in
the stroma, particularly in S-G2-D2. In
addition to forming such a large part of
the corneal soluble protein, it was one
of only two intrinsic (i.e., nonserum) cor-
neal proteins which were sufficiently antigenic to stimulate antibody production in
rabbits immunized against unfractionated
corneal proteins. At present all that is
known about this protein fraction is that
it has a "beta" mobility on polyacrylamide
gel electrophoresis, elutes from DEAEcellulose at about 0.045 M, and is not a
serum protein, glycoprotein, lipoprotein, or
a soluble collagen. Other epithelial proteins are present in smaller amounts. Of
these, the four or five in E-G2-D1 constitute 17 per cent of E-G2, whereas the other
fractions are mainly in trace amounts. The
lack of detectable serum proteins in the
epithelium was consistent with the findings
of other studies.3'5
The present study documents the existence of serum albumin and 7S-gamma
globulin in the stroma by the use of specific
antisera against these two proteins and relates these antigens to certain disc electrophoretic zones. The identity of the S-G2D2 serum protein as serum transferrin rests
on the specific staining of the transferriniron complex by Nitrosos-R salt14 plus the
characteristic disc electrophoretic patterns.24 The presence of a serum lipoprotein is suggested by the correlation of
a serum antigen of beta mobility with the
disc electrophoretic zone Rf 0.08 in S-Gl
which stains with oil Red-O.14 There is
also present in S-Gl a fifth serum antigen
of slow beta mobility, the identity of which
was not demonstrated. The suspicion that
it was haptoglobin could not be documented by benzidine15 staining.
The four serum proteins demonstrated
make up 3.3 mg. of the 5.9 mg. (56 per
cent) of soluble protein in the stroma, the
remaining 44 per cent being made up of
proteins intrinsic to the stroma and the
unidentified serum protein. Expressed differently, of the 2.8 per cent of the stromal
wet weight which is soluble protein, 1.2
per cent is stromal-specific protein and
1.6 per cent is of serum origin. The serum
proteins present in the stroma are those
with the highest concentrations in blood
except for the 19S and lipidic macroglobu-
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
Soluble proteins of bovine cornea 125
Volume 12
Number 2
lins,22- -3 which are too large to enter the
cornea by diffusion.7 If serum proteins enter
the cornea by diffusion from the limbal
capillaries, then their comeal concentrations should be proportional to those in
serum. A ratio of the absolute values of
the serum concentrations of each of these
four proteins in grams per 100 ml. as given
in a standard text23 to the stromal concentrations in milligrams per stroma as found
in the present study, gives the following
values: albumin 4 to 6, 7S-gamma globulin
0.5 to 1.0, transferrin 1.0, and lipoprotein
0.75 to 1.0. A ratio of about 1.0 appears
to hold except for albumin, which is present in the cornea in much smaller proportion than the other three. This can be
explained by assuming a diffusional loss of
albumin into the aqueous as suggested by
the calculations of Maurice and Watson.7
The value of 0.7 mg. albumin per stroma
converts to roughly 0.3 Gm. per 100 ml.
stroma, about 1/10 the average serum
value, a figure which is consistent with
Maurice and Watson's predictions.7 The
estimate of 1.8 mg. gamma globulin per
calf cornea is consistent with Stock and
Aronson's4 figure of 0.5 mg. per human
cornea.
The nonserum proteins found in the stroma comprise 44 per cent of the soluble
stromal proteins. None are collagen. Certain proteins are present in fairly large
amounts. Of these, the Rf 0.6 to 0.7 zone
of S-G2-D3 stains strongly with PAS and
must have a high carbohydrate content.
Also in this fraction are a number of more
faintly staining glycoprotein zones. Perhaps some of these correspond to the soluble neutral and acid mucopolysaccharides
described by Dohlman and Balazs0 in
1957. The series of repeating electrophoretic zones in S-G2-D4 may represent a
stromal protein or proteins made up of
numbers of basic units in varying degrees
of aggregation, perhaps analogous to the
human haptoglobin types 2-1 and 2-2 seen
in disc and starch gel electrophoresis.15'21
The possible relation of the several serum, stromal, and epithelial soluble pro-
teins to the problem of host sensitization
to donor material in keratoplasty is an
area which is in need of exploration. If,
in fact, soluble comeal components are
important in graft rejection, then such techniques as prior soaking of the graft and
transplantation antigen typing would assume more importance.
REFERENCES
1. Maurice, D. M., and Riley, M. V.: The Cornea. In Graymore, C. N., editor, Biochemistry
of the Eye, New York, 1970, Academic Press,
Inc. pp. 76-86.
2. Polatnick, J., Tessa, A. J., and Katzin, H. M.:
Comparisons of collagen preparations from
beef cornea and sclera, Biochim. Biophys.
Acta 26: 365, 1957.
3. Kawerau, E., and Ott, H.: The soluble proteins of the cornea, paper, and immunoelectrophoretic studies, Exp. Eye Res. 1: 137,
1961.
4. Stock, E. L., and Aronson, S. B.: Corneal immune globulin distribution, Arch. Ophthalmol.
84: 355, 1970.
5. Allansmith, M., Newman, L., and Whitney,
C : The distribution of immunoglobulin in
the rabbit eye, Arch. Ophthalmol. 86: 60,
1971.
6. Dohlman, C. H., and Balazs, E. A.: Carbohydrate-containing protein fractions in the corneal stroma, Acta Ophthalmol. 35: 454, 1957.
7. Maurice, D. M., and Watson, P. C : The
distribution and movement of serum albumin
in the cornea, Exp. Eye Res. 4: 355, 1965.
8. Elliott, J.: Immune factors in corneal graft
rejection,
9.
10.
11.
12.
13.
14.
INVEST.
OPHTHALMOL. 10:
217,
1971.
Leibowitz, H. M., and Luzzio, A. J.: Transplantation antigens in keratoplasty. I. Isolation of a soluble fraction with transplantation
antigen activity, Arch. Ophthalmol. 83: 215,
1970.
Warburg, O., and Christian, W.: In Colowick,
S. P., and Caplan, N. O., editors: Methods of
Enzymology, New York, 1957, Academic
Press, Inc., Vol. Ill, p. 73.
Lowry, O. H., Rosebrough, N. J., Farr, A. L.,
et al.: Protein measurement with the phenol
reagent, J. Biol. Chem. 193: 265, 1951.
Davis, B. J.: Disc electrophoresis. II. Method
and application to human serum protein, Ann.
N. Y. Acad. Sci. 121: 404, 1964.
Zacharius, R. M., Zell, T. E., Morrison, J. H.,
et al.: Glycoprotein staining following electrophoresis on acrylamide gels, Anal. Biochem.
30: 148, 1969.
Beaton, G. H., Selby, A. E., and Wright, A.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017
126
15.
16.
17.
18.
19.
20.
Investigative Ophthalmology
February 1973
Holt and Kinoshita
M.: Starch gel electrophoresis of rat serum
proteins, J. Biol. Chem. 236: 2001, 1961.
Smithies, O.: Zone electrophoresis in starch
gels, Adv. Prot. Chem. 14: 65, 1959.
Schneider, W. C : In Colowick, S. P., and
Kaplan, N. O.: Methods of Enzymology. New
York, 1957, Academic Press, Inc., Vol. Ill, p.
680.
Neuman, R. E., and Logan, M. A.: Determination of hydroxyproline, J. Biol. Chem.
184: 299, 1950.
Prockop, D. J., and Udenfriend, S.: A specific
method for the analysis of hydroxyproline in
tissues and urine, Anal. Biochem. 1:228, 1960.
Ouchterlony, O.: Antigen-antibody reactions
in gels, Acta. Pathol. Microbiol. Scand. 32:
231, 1953.
Scheidegger, J. J.: Une micro-methode de
l'immuno-electrophorese, Intern. Arch. Allergy
7: 103, 1955.
21. Pastewka, J. V., Ness, A. T., and Peacock,
A. C : Electrophoretic patterns of normal
human serums by disc electrophoresis in
polyacrylamide gel, Clin, Chim. Acta 14: 219,
1966.
22. Gerbarg, D. S.: The clinical significance of
serum glycoproteins, in Sunderman, F. W.,
and Sunderman, F. W., Jr., editors, Serum
Proteins and the Dysproteinemias. Philadelphia, 1964, J. B. Lippincott Co., p. 185 and
186.
23. White, A., Handler, P., and Smith, E.: Principles of Biochemistry, New York, 1968, McGraw-Hill Book Co., p. 705.
24. Kristjansson, F. K., and Hickman, C. G.:
Transferrin alleles in cattle, Genetics 52: 627,
1965.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/932875/ on 06/14/2017